Effectiveness of Olutasidenib Versus Ivosidenib in Patients with Mutated Isocitrate Dehydrogenase 1 Acute Myeloid Leukemia Who Are Relapsed or Refractory to Venetoclax: The 2102-HEM-101 Trial Versus a US Electronic Health Record-Based External Control Arm

Introduction

Combination therapies involving venetoclax (VEN) are common treatments for acute myeloid leukemia (AML) in patients ineligible for intensive chemotherapy. Although patients on a first-line VEN-based regimen may achieve a response, most patients relapse or become refractory to treatment (R/R). Isocitrate dehydrogenase 1 (IDH1) inhibitors olutasidenib (OLU) and ivosidenib (IVO) are both approved treatments for patients with R/R AML and an IDH1 mutation (mIDH1), but limited data is available to inform the optimal sequencing of therapies. This study aimed to provide the first evidence on the comparative effectiveness of OLU versus IVO in mIDH1 R/R AML.

Methods

Outcomes of OLU-treated patients from the 2102-HEM-101 single-arm trial were compared with a real-world external control arm (RW-ECA) of IVO-treated patients, constructed from the Loopback Analytics electronic health record database. Target trial emulation methodology was used to derive the RW-ECA by applying eligibility criteria of the 2102-HEM-101 trial. Eligibility criteria were prioritized based on their prognostic value through consultation with clinical experts. Recognizing a trade-off between bias and sample size, the first analysis (analysis 1) applied only criteria related to age and central nervous system or other tumor location involvement, e.g., spinal cord compression, while a second (analysis 2) additionally applied all remaining criteria where ≤3 patients failed to meet each of those criteria, including liver and renal function, performance status, uncontrolled infections and metabolic disorders, and prior solid organ allograft, stem-cell transplant, anticancer therapies and malignancies. Due to longer follow-up among OLU patients, a sensitivity analysis was conducted where patients in the OLU-arm were censored at the maximum follow-up observed in the RW-ECA. Effectiveness outcomes included composite complete remission (CRc) as best response (comprising complete remission [CR], CR with partial hematological recovery [CRh], CR with incomplete recovery [CRi]), overall survival (OS), and transfusion independence (TI).

Logistic regression was used to compare binary outcomes reporting the odds ratio (OR) and 95% confidence interval (CI). Kaplan–Meier approach was used for survival outcomes.

Results

Twelve of 153 patients in the 2102-HEM-101 trial previously treated with a VEN-based regimen. 20/108 patients treated with IVO after VEN in the Loopback dataset contributed to analysis 1, and 14 patients contributed to analysis 2. 75% and 80% of patients met at least 80% of all eligibility criteria in analysis 1 and 2, respectively. Median time on prior VEN for OLU- and IVO-treated patients was 6.75 and 5.24 months respectively.

Rates of CRc as best response were higher amongst OLU-treated patients (6/12 comprising 3 CR,1 CRh and 2 CRi; 50%) vs IVO-treated patients (4/20 comprising 0 CR, 3 CRh and 1 CRi; 20%) in analysis 1 (OR 4.00 (95% CI, 0.86 – 21.05)), with statistically significant differences observed in analysis 2 (50% vs 14.3%; OR 6.00 (95% CI, 1.03 – 50.46)). Median OS was 16.23 months for OLU-treated patients compared with 5.45 months and 4.57 months for IVO-treated patients in analyses 1 and 2, respectively. The proportion of patients surviving 6, 12, 18 and 24 months was numerically higher amongst OLU-treated patients. At 24 months in analysis 1, 42% (95% CI, 20% - 87%) of OLU-treated patients had survived compared with 22% (95% CI, 8.8% - 57%) of IVO-treated patients. No IVO-treated patients survived 18 months in analysis 2. OLU-treated patients also had higher odds of TI compared to IVO-treated patients in analysis 1 (OR 6.37 (95% CI, 0.69 – 140.71)) and analysis 2 (OR 4.50 (95% CI, 0.48 – 100.42)) although not statistically significant.

After censoring OLU-treated patients at the maximum follow-up of the RW-ECA, the differences in outcomes between OLU- and IVO-treated patients were similar to those observed in the main analysis.

Conclusion

This study provides evidence suggestive of favorable effectiveness of OLU versus IVO for patients with mIDH1 AML who are R/R to a VEN-based regimen. Limited sample sizes in this rare indication may explain the common absence of statistically significant differences. As with any RW-ECA study, limitations include the risk of unmeasured confounding and differential outcome reporting.